[ILUG] Citrix ICA cliient installation

[Eoin Kelly]

Hi Damien,

There should be a file in the top level directory of what you extracted
called README and/or  INSTALL-these should give you the details you need.
However...

 If it's a binaries of the client then there'll probably be an installation
script included that will copy the files to the correct locations. Open your
favourite terminal(konsole?) and change to the directory that you extracted
the .tar.gz file to and type "sh NAME_OF_INSTALL_SCRIPT" - the README or
INSTALL files should tell you what the script is called.

If what you've downloaded is source code[0] then (in a terminal)change to
the directory you extracted the tar.gz file to and do the following:

1) The source code must be set up for your particular system. This is done
by running a configure script(it sets certain variables, checks to see if
you have all the stuff you'll need to actually compile the program, sets the
directory that the binaries will be installed to etc.). You can invoke this
script by typing "./configure" (without the inverted commas but *definitely*
with the dot-forwardslash [1]) on the command line. you can pass various
options to this script to change for example where you want the program
installed to. Typing "./configure --help" will give you a full list of these
options. By default gnu software seems to install  under /usr but this may
not be the same in your case(the output of "./configure --help will tell you
what the default is).If there's other stuff you need to install configure
will tell you so have a careful read of it's output.

2) The next step is to actually compile everything. This is done by invoking
the Make utility which will compile everyting in the correct order and sort
out all the dependencys for you etc. --simply type "make" on the command
line and sit back-don't worry too much about the stuff that's printed on
your terminal screen by Make-it's just the output of the compiler(amongst
other things) and if you sorted out any problems reported by the configure
script then you shouldn't have problems here(unless you run out of disk
space like me ;)...think of it like that falling code in the matrix.... Once
make is finished you'll get the command prompt back on your terminal and
you'll be ready for step 3...

3) So far nothing  has been installed or changed on your system. All the
files that Configure and Make create are placed under the hierarchy that was
created when you extracted the tar.gz file (i.e. under your current working
directory). You must now install these newly created binaries as well as the
necessary configuration and documentation files i.e. copy them to where
linux and other programs that may use them  expect to find them. To do this
type "make install" from the command line. This will copy all the necessary
files to where linux can find them easily and also takes care of setting the
permissions too.

A useful exercise might be to do a test run-create a new directory on your
desktop (I'll call it mytestdir). When you run configure tell it that you
want to install the software to this directory-do this by typing
"./configure --prefix=/home/YOUR_LOGIN_NAME/Desktop/mytestdir" (replace
YOUR_LOGIN_NAME with the name you use to login)[2]. This means that when you
go through the make and make install steps above the files will be installed
to "mytestdir" instead of their usual place. NOTE: DO NOT DO IT THIS WAY
WHEN YOU ACTUALLY WANT TO INSTALL PROPERLY-you can now have a look at your
handywork under mytestdir. When you're feeling confident with the whole
process delete mytestdir and re-run configure,make,make install (as in steps
1->3 above) this time specifying where you *really* want the files to be
installed. Before you re-run configure you should probably delete a file
called configure.cache in the directory that you extracted the tar.gz
to-this file is a cache (surprise surprise..) of settings already detected
by the configure script-it mightn't matter but then again...best to delete
it and let configure start over.

If you're feeling brave all of the above can be truncated into typing
"./configure && make && make install" on the command line. The && don't
allow the command following  to run unless the previous one was successful.


[0] If you've downloaded the source code you'll need a "development
environment" in order to build the software i.e. you'll need  _at_least_
the gcc compliler and two packages called "autoconf" and "automake" as well
as some header files - you'll find automake and autoconf on the SuSE CDs but
most of what you need gets installed by default. You'll probably need other
stuff but the output of the configure script will tell you if you're missing
anything else.

[1] When you type a command in linux it only searches certain directories to
find the executable. The directory you created when you extracted the tar.gz
will not be one of them so when you run the configure script you have to
tell linux exactly where this script is-otherwise it'll try the usual
spots,fail and complain to you about it. A single dot is shell shorthand for
your current working directory so by typing "./configure" you're telling
linux "run a program called configure which is located in the directory I'm
in at the moment". You'll probably have noticed that you don't have to do
this for "make"- This is because the make program(which comes in the
automake package) is actually in one of the directories that Linux searches
and is smart enough to figure out that is should look for input files in the
directory you're currently in.

[2] I've had problems using the ~ shortcut for my home directory in these
situtations so it's probably best to specify the path fully.

Whew!
HTH,
-Eoin